Acer 6920 Laptop CPU overheating issue

Hey everybody,

I have been searching through all the threads posted here about Acer 6920 Laptop, and Acers in general and have found none than speak to my specific problem.

Recently, my CPU has started heating up almost instantaneously to ridiculous temperatures (100 degrees + via RealTemp) This overheating occurs in any flash based video program (ie Youtube/Hulu) and in most every game I can play on my laptop.

The first few seconds after the game or movie starts, CPU temperature spikes from an idling 35-40 degrees to 96-100+ degrees within 3-4 seconds. This is all at the maximum CPU multiplier of 12x.

These are the issues:

At high processor loads (where the multiplier is 11x, 12x and 13x with IDE) the temperature of both CPU cores rapidly rises from around 40 (44) degrees C idle to 97 (103) degrees idle within 5-10 seconds of the load being applied.

The temperature values are taken from HWMonitor and Realtemp, with the RealTempt value in brackets.

This overheating results in throttling, where the CPU cycles from 12x or 11x speed (2.4 and 2.2 GHz approximately) down to 6x (1.2 Ghz) rapidly. The throttling seems to stabilise the CPU temperature at around 97 (103) degrees C, about 3 degrees below the shutdown threshold in RealTemp.

In every possible torture test and real world usage scenario, the laptop automatically shuts down (ie, all power is lost instantly, equivalent to holding down the power button) in almost the EXACT SAME TIME PERIOD : 10 min 30 seconds to 10 minutes 45 seconds. I repeated the Orthos "small FFTs-stress CPU" test around 15 times, along with tests using full screen youtube videos, flash videos and video games.

Stuff I have tried already (which is not to say I may have done things wrong)

Software

Formatted all drives multiple times
Reinstalled Windows 7 professional 64x multiple times
Ran tests in Safe Mode
Flashed BIOS from version 1.07 to 1.16
Updated to Nvidia 9500GS drivers 197.16

Hardware

Cleaned Dust from back panel mesh covering the vents
Cleaned Dust from fan
Cleaned dust from heat sinks
Tightened Heat Sink screws

Undervolting

Undervolted using RMClock to minimal values for 12x, 11x, 10x, 9x speeds. (1.1375V, 1.0875V, 1.075V, 1.075V respectively)

12x speed at any voltage resulted in near instantaneous rise of temperature to 97 (103) at which point it held steady until the 10:40 mark, at which it invariably resulted in shut down

11x speed at any voltage resulted in a SLOWER rise of temperature but still ended up at 97 (103) after ~ 8 minutes, again the laptop shutdown at the 10:40 mark

10x speed at any voltage resulted in a MUCH SLOWER rise of temperature to around 76 (90) after 10 minutes, but yet again the laptop shut down at the 10:40 mark

9x speed at any voltage resulted in the SLOWEST rise of temperature to around 73 (80) . I have not yet checked to see if it shuts down

THE ONLY WAY to stop the laptop shutting down for any long period of time under stress is to throttle it by 50%. In this case the temperatures stay below 70 (80) and the computer does not shut down, However performance is severely compromised as the processor is running at an effective 1.2 Ghz.

TLDR: My CPU overheats like crazy, I have tried undervolting, formatting, flashing BIOS, and dusting. Help?

 

Welcome to the forums Kasseev; I wish it was under better circumstances.
Although I must admit, it's almost refreshing to see a case of an Aspire 6920G actually overheating and throttling.

That kind of rapid heat gain shouts one of two possibilities - a lack of connection between the CPU and the heatsink, or a malfunctioning fan.

We can eliminate the latter possibility if the GPU temperatures are sane; considering the two share a common copper heatpipe to the fan, high GPU temperatures would point to a fan failure. Conversely, regular GPU temps would point to the CPU not having good/any contact with the heatsink.

When you were cleaning things out, did you accidentally scrape off the thermal compound that connects the CPU to the heatsink? Is the heatsink screwed back on properly?

You need to take the fan out to clean the copper grille it blows on. Did you reinsert it properly? Did you reconnect it to the motherboard?

 

I would definitely look at reapplication of a thermal compound on your CPU...

 

lol, wow, kind of a no brainer there. I guess I didn't want to consider swapping out the compound because I have never done something like that before.

In any case, thanks a lot for the quick replies,
Amazon.com: ArctiClean 60ml Kit (includes 30ml ArctiClean 1 and 30ml ArctiClean 2) and 3.5grams Arctic Silver 5 Thermal…

is in the mail as I speak.

Hopefully this will solve the issue.

 

Let us know if you need any help with thermal compound application. Either way, just remember that too much is just as detrimental as too little. Hope things get resolved!

 

Let us know how things go, too, and be sure to stick around afterward.
NBR's a pretty spiffy place.

 

Hey again,

So the Arctic Silver showed up today, along with a nifty new tool kit.

Attached is the thermal compound replacement job I did, it would be cool if you guys could tell me if thats ok or not.

Im benchmarking stuff as I type, and so far besides from a few screw ups it has gone fine.

Using the orthos torture test again the temperature is rising far more slowly, and it is currently at holding steady at 44-46 degrees after five minutes with the use of an NZXT Cryo LX.

There was one test run immediately after I added new compound where it slowly rose all the way to 100 degrees again over ten minutes. I went back in and added more compound (the photo was taken after this) and now it seems to be better but time will tell.

Edit: lol didn't realise these boards were censored :O

If the worst should happen, is there any other explanation for this behaviour?

 

Thermal compounds increase in efficacy after a few thermal cycles. After about 200 on-off cycles, your AS5 should be working at top cooling.
If it relapses, I wouldn't know what the problem is.

The application on the CPU looks good - you want enough to cover the whole thing, but not too thick.
However, I wouldn't have put it on the northbridge - that's the larger of the two chips. The heatsink should have a heatpad for that one, and adding thermal compound to a heatpad can sabotage its efficacy.

 

Uhh, well I removed the heatpad entirely, assuming it wasnt doing its job right considering that temperatures always hit the roof.

In any case temperatures seem back to normal, CPU is not going above 46 degrees no matter what I do to it

Now that it is fixed, I realise how slow the computer was in the first place haha, any tips on if this can be overclocked?

 

The northbridge doesn't have a thermal sensor like the CPU does, so you had no real way of determining whether or not it was overheating or not.
In most cases, the northbridge isn't the problem.
Then again, if the application of the thermal compound was good, it won't make a difference.

Overclocking the CPU? No. You can replace it though - I did!
Overclocking the GPU? Yep! Just get the latest driver straight from Nvidia.com, and then install Nvidia System Tools or RivaTuner.

 

Oh wow ok,

So wait, what does it mean then when it gives me temperature values for two cores?

And is it worth buying a new CPU for the 6920 or should I just get a new laptop? It seems to be fairly old at this point, starting to have trouble with most new games.

 

Well, that's because your CPU has two cores; they're on the same piece of silicon, right beside each other.
They each have their own thermal diode, but they rarely vary too much between the two; for me, one core is permanently 1-2 degrees higher than the other.

Then by all means get a new laptop! Some of the new Acers have Radeon HD 5650s in them, and those will eat the 6920G's GeForce 9500M GS for lunch.
I'm personally going to run out the extended warranty on mine, then keep it as a media center/Blu-ray player.